# tx

> Primitives, not frameworks. Headless infrastructure for memory, tasks, and orchestration.

tx provides composable primitives for agent infrastructure — task management, memory, coordination, and observability. You own the orchestration; tx owns the plumbing.

## Docs

- [Overview](https://tx-docs.vercel.app/docs): What tx is and why primitives beat frameworks
- [Getting Started](https://tx-docs.vercel.app/docs/getting-started): Install tx and run your first agent loop

## Headful Experience (Dashboard)

- [Headful Experience](https://tx-docs.vercel.app/docs/headful): Dashboard UX overview
- [Tasks and Sub-Tasks](https://tx-docs.vercel.app/docs/headful/tasks-and-subtasks): Task list, task detail, and child task bulk operations
- [Command Palette](https://tx-docs.vercel.app/docs/headful/command-palette): `Cmd+K` / `Ctrl+K` global actions and navigation
- [Docs, Runs, and Cycles](https://tx-docs.vercel.app/docs/headful/docs-runs-cycles): Non-task dashboard surfaces and empty states

## Primitives

- [tx ready](https://tx-docs.vercel.app/docs/primitives/ready): Get next workable task (unblocked, highest priority)
- [tx done](https://tx-docs.vercel.app/docs/primitives/done): Complete task, potentially unblocking others
- [tx block](https://tx-docs.vercel.app/docs/primitives/block): Declare dependencies between tasks
- [tx claim](https://tx-docs.vercel.app/docs/primitives/claim): Claim task with lease to prevent parallel collisions
- [tx memory](https://tx-docs.vercel.app/docs/primitives/memory): Search, navigate, and manage your markdown knowledge graph
- [tx pin](https://tx-docs.vercel.app/docs/primitives/pin): Persist named context blocks for reuse across agents
- [tx gate](https://tx-docs.vercel.app/docs/primitives/gate): Add human approval checkpoints between phases
- [tx guard](https://tx-docs.vercel.app/docs/primitives/guard): Set task creation limits to bound agent autonomy
- [tx verify](https://tx-docs.vercel.app/docs/primitives/verify): Attach machine-checkable done criteria to tasks
- [tx reflect](https://tx-docs.vercel.app/docs/primitives/reflect): Run session retrospectives and tune agent behavior
- [tx label](https://tx-docs.vercel.app/docs/primitives/label): Organize tasks with labels for ready queue scoping
- [tx sync](https://tx-docs.vercel.app/docs/primitives/sync): Persist to git-friendly JSONL
- [tx send / tx inbox](https://tx-docs.vercel.app/docs/primitives/inbox): Send and read channel-based agent messages
- [tx trace](https://tx-docs.vercel.app/docs/primitives/traces): Inspect execution traces for agent runs
- [tx doc](https://tx-docs.vercel.app/docs/primitives/docs): Docs-as-primitives with YAML structured documentation
- [tx decision](https://tx-docs.vercel.app/docs/primitives/decision): Decisions as first-class artifacts with review lifecycle
- [tx spec](https://tx-docs.vercel.app/docs/primitives/spec-trace): Spec-to-test traceability and FCI phase scoring
- [tx invariant](https://tx-docs.vercel.app/docs/primitives/invariants): Doc-derived invariants with enforcement and check recording

## Interfaces

- **CLI**: Scripts, terminal workflows, agent loops
- **MCP Server**: Claude Code integration (42 tools)
- **REST API**: Custom dashboards, external integrations
- **TypeScript SDK**: Programmatic access from your agents
- **Dashboard**: Visual monitoring and management

## Quick Start

```bash
curl -fsSL https://raw.githubusercontent.com/jamesaphoenix/tx/main/install.sh | sh
cd your-project
tx init
tx add "My first task"
tx ready
```

## Optional

- [GitHub](https://github.com/jamesaphoenix/tx)